e-mini daily profit target

Agreed. But more data is not worse than less
You need the rules of the system to take advantage of the underlying workings of the market that are present day in and day out and more data will help you rule out the bad ideas that are masked as good ideas when only using 2months of data.

 

If it's an intraday system, you can backtest (and cautiously optimize) it on randomly selected months from the past few years. Use the other months for out of sample testing.

I would concur with what others have said about 2.5 points/contract/day - it is ample profit.

 

But if he can do it every day its awesome. He can always increase size.

Nitro, you do pretty big volume even though its based on singles and doubles. Do you feel commission eats up the profits a bit?

I ask cause I try to trade that way occaisionally, but can't seem to make it work. All the power to you with the approach though.

Jay

 

Averaged daily profit for two months really doesn't tell you much. What is average profit per trade? If it's $5 and you are doing 30 trades a day, it is far different from one trade a day or one trade every three days.

Two months is not enough backtesting. If you are experienced, you might be willing to try it real time based on that, because you would have an idea of what works and what doesn't. For a newbie, that approach could be deadly. I have seen many daytrading systems that worked great for a couple of months but sucked over the long haul.

I don't really understand the objection to doing backtesting. My view is more data is better than less. If a system didn't work over long periods of time in the past, that surely warrants caution going forward. Of course, it is possible to curve fit using historical data, but the more out of sample testing you do, the less likely you have curve-fit data.

I think it is important to use data from different market environments. If I were backtesting an S&P system, I would want to use data from the '90-'95 time frame because that was a tough period. Bubble market data is not that useful in my opinion unless you think we will encounter another bubble market soon. You might have to normalize parameter values but that is pretty simple to do.

I wouldn't turn up my nose at a system that produced 2.5 ES points per day, particularly since there probably are very few people trading the ES who are profitable at all.

 

1 - 2 points a day (average, and net of commissions and all trading costs) is great... just gradually increase your size to 10 contracts, and this is $500 to $1k a day average...

 

More data is better than less, IF the data is still relevant. The argument could be made that the market has changed enough in the last couple of years that the data from that long ago may not produce valid backtesting results. I would definitely at least weigh the results from the last few months more than the results from a year or two ago. Any system that is tuned to what the market is doing now instead of set up to work in any market should have better results, as long as that system is continuously tuned.